Reconsider Problem $3\mathrm{.}17.$ Partition the $4,985$ historical records of interest into a training partition $(60$ percent of the records$)$ and a validation partition $($the remaining $40$ percent of the records$).$

Determine the accuracy, sensitivity$,$ and specificity for the KNN model with

$$

$=$

$10$

$($based on the historical data in the training partition$),$ when it is used to classify applicants as to whether they are likely to default $($defined as greater than a $10$ percent chance of default$).$

Repeat part a with the classification tree model, with a maximum of four splits as the stopping rule and using the full tree.

Repeat part a with the na$$ve Bayes model using the data on the Binned Data worksheet tab.

Repeat part a with the logistic regression model.

Comment on which model$($s$)$ perform best.